Bathymetry Mapping using Landsat ETM+ Data and Field Measurements for West Coast of Yemen

Abstract

The bathymetry of coastal waters in the Red Sea coastal water is of the vital importance for shipping safety because of presence of navigational hazards. We used remote sensing data from Landsat-7 (ETM+) for bathymetric mapping in Al-Luhaia port, Western Yemen. We used a global positioning system to locate the accurate sampling points for sea depth. An echo sounder was used to collect sea depth information. We examined suitability of wavelength bands for bathymetry. This paper puts forward a method to extract water depth information from multispectral data Landsat-7 (ETM+). We applied simple linear regression to relate field measured water depths to pixel brightness values in the blue band of Landsat-7 (ETM+) multispectral imagery that had been corrected to at-satellite reflectance using published calibration coefficients. The regression relationship at the clear shallow water site was accurate (R2 = 94.40% band 1) for water depths in the range 2 to 9m. The regression analysis offered by the model has been verified using data from 1500 points (depths) that were not used in model generation, were used for testing the validity. This validation test model offered a very good correlation coefficient, 0.9385. We made a comparison between actual depth measured by hydrographic echo sounder during the field measurement and the estimated depth derived from satellite images to establish the margins of error in the estimates. A mean of error of 4.08%, with an accuracy of 95.92% was found.